801D4 measurements

[boomerang]

Hi everyone, let’s start with the best regards from Holland.
First posting here, read a lot already, great forum!

Question:

In the last week, I did some measurements on my 801D4.
Glad to obtain a frequency respons quite similar to the results of HifiNews in their review.
What wonders me is why /how BW changed the output of the woofers between 90 en 400 Hz, (minus 2db) also compared to the 800D3. (As HifiNews also mentioned)
I measured the output of the low pass crossover and there could be a kind of explanation.
The transfer curve is not what I expected from a 18db Butterworth filter.
It has some bumps in its characterisic.
So maybe it isn’t Butterworth?
Coil and capacitor values also point away from a straight butterwort design.
There is a +2 dB ripple at 70hz, maybe a measuring faillure, maybe the answer?
Repeated this measurement several times, always the same result.
Btw: all of this measuring started because I wanted to try Bi amping, leaving the original cross over for mid/high (absolute top components here) but making low pass analog active, still third order at 420hz. (Since level matching will also be necessary)
But surely, that 2db ripple at 80 hz will be gone then, so what will happen with sound quality....
Exchanged MC312 for 2 x Michi M8 a month ago.
MC312 can not cope with Michi regarding low frequency power and control, but I think the Mcintosh was a little more refined in the mids/high.
So bi amping with both of them, could maybe make me even a little more happy.
Or maybe not at all...because of crossover challenges.
Should there be someone here who also did some measurements and/or knows more about the crossover layout of the 801D4, glad to hear your comments!

Update: did a verification measurement tonight on a second order passive Butterworth crossover, loaded with a test resistor of 10 Ohm.
Curve was exactly what could be expected, no ripples, -3db spot on.
So the measurement is/was valid!
Could BW deliberately have altered component values, to obtain a certain non linearity from the crossover, in order to emphasize a certain part of the frequency band?
Low pass components: From plus terminal, first coil = 2.7mH, then a set of capacitors (total 147uF) to negative terminal, followed by a 1mH coil connected to the positive terminal of the woofers.
A plain third order low pass, but with deviating component values.
Who can shed a light on this?
And am I right to presume that exchanging the passive low pass for a plain Butterworth active one, will make me lose some output in the lower frequency band in this case? (40-80Hz)

 

[Livarot]

Hello Boomerang,

It looks like your measurement process & results are correct (but a schematic with measurement points would be the 100% foolproof way to confirm).
Your results deviate from a textbook filter response for the following reasons:

• Woofers in a vented box have a rather complex impedance curve which is VERY different from a perfect resistor. The impedance has too peaks on both sides of the vented box tuning frequency (Fb), the impedance magnitude at these peak is much largest than your DCR impedance. I've included a typical example from Leach's classic book (amazon link )

• 
  Because of the above impedance, your passive filter doesn't exhibit a classic response (the filter response is affected by the impedance it sees). It is also possible that B&W chose a non-flat filter response near the cut-off frequency (intentional ringing around 300 Hz)
• In short, the non-perfect filter response is a consequence of both the woofer electrical impedance & B&W design choices

I hope this clarifies matters!

 

[Livarot]

Hello Boomerang,

Some additional comments about your bi-amping idea:

• Because of the passive crossover LF response, bi-amping with an active filter (using "standard low pass filters") will results in a different frequency response, unless you try to match the passive crossover response ripples with added parametric EQ sections (possible but tedious to tune)
• For the above reason, you might no like the end result using a standard low pass !
• It might therefore be safer to keep the passive crossovers even when bi-amping
• You should also make sure that both amplifiers have the same gain if you take the bi-amping route (sorry if this is obvious to you!). I wonder if your initial listening was affected by gain differences (the gain should be matched within 0.1dB)

 

[boomerang]

Thanks Livarot!

Exactly as i was thinking tonight....best thoughts tend to pop up just before going to sleep
I already had planned a crossover measurement with a plain resistor instead of the woofers for today.
But it is clear to me now, that this is useless.
Question that remains is this: will the 2 db output drop at 75 hz (black curve) cause a problem, going active?
I presume that the flat active crossover output, will also flatten the woofer output more in the direction of the former D3. (red line)

This could be an advantage, but maybe not.
Will have to try and listen.
Last question;
When trying to find the electrical -3db crossoverpoint with a scope, i found this picture.
Will the phase respons at cross over alter noticable, when we go from a "divergent passive Butterworth" to a perfect active Butterworth?

As you can see, output at 420hz is not a 100% shift of 90 degrees (270 total), but more like 75 degrees (255)
All electric of course.
Acoustic phase is very hard to determine accurately with my limited equipment and knowledge.

 

[Pearljam5000]

Measurements aside, how's the sound?

 

[boomerang]

Ah, how is the sound...indeed the most relevant question to ask.
Came from CM10, could not believe that the new cone material continuum was relevant enough to go to 702S2, till I heard the difference.
In 2020 a dream came true.
702S2 out, 802D3 in.
Great live-like sound experience, transparant, detailed, smooth.
A huge jump in sound quality.
Till then, I more or less listened to speakers.
Now I was suddenly listening to music! (While 702S2 really is a great sounding speaker allready)
Did not know meanwhile, that the D4 series was already announced.
Was able to sell my one year old 802’s for a good price and ordered the 801D4.
Had serious doubts though...802 or 801.
First impression; did I maybe overdo it with the 801 in my average sized room?
They really seem to deliver the lowest frequencies with so much ease, yet fast and very powerful, its almost bizarre.
But also, an even bigger live presentation, more open midrange, yet a more refined high.
And in my room, a smaller sweet spot, due to room nodes that are more pronounced now.
Already mounted huge bass traps in four corners, better now, but still not perfect.
The Michi M8 mono’s made the bass tighter and even faster, but bas weight increased as well.
At the end of this story, I tend to think that I would have been just as happy with the 802 running on the Mcintosh MC312.
On the other hand, then I would have always been busy asking myself ”shouldn’t I have bought the 801”

 

[boomerang]

The active copy of the passive 18dB/okt low pass

The low pass frequency response measured,

The passive and active acoustic measurement, green passive, yellow active.

The 80hz bulge that is measured in most of the reviews, is tamed now.
Bridging the passive low pass, means the 0.4 ohm series resistance caused by the coils, is deleted now.
How this influences low frequency response, I can not track down, but the overall low frequency performance has improved in my opinion.
The low end sounds tighter and a little less “fat”, cleaner,... hard to describe in fact.

 

[Livarot]

Hello boomerang,

Thanks for sharing the update, interesting results.
As you've bypassed the passive crossover (and therefore the DC resistance of the inductors), you've effectively increased you LF output a little (typically 0.5 to 1dB, this can be easily calculated knowing the woofer impedance curve + crossover schematic) but reduced your LF impedance (by 0.4 ohm) so you are drawing more current from the amp for the same output voltage.

Good to know that you like the end result! This might be quite room dependent as well (a little bump at 80Hz might be beneficial in some rooms and a problem in others). I've got 800D3s but I am using DIRAC with a miniDSP SHD.

Last idea, you could try to add another low pass around 1 KHz. This will not impact the system response (as the woofer is already -25dB down at 1000Hz) but will attenuate further the woofer response above 1 KHz (where the woofer mechanical resonances/break-ups will be), you might get a slightly cleaner response (a blind A/B will confirm if this is audible or not!!).

 

[boomerang]

Hi Livarot,

Thanks for your response!
When matching the outputlevels of the amps, i reduced the bass amp a tiny bit below the mid/high amp because of that 0.4 ohm.
My worries where more about bass reflex tuning.
In the more serious bass reflex online calculators, you can add a series restistance (from cables/cross over)
This influences the total Qts (woofer) and so port tuning and enclosure volume.
But they stay the same of course.
I presume F3 shifts slightly upwards now and total Q will be slightly lower.
But again, after listening a lot of well known tracks again last evening, i am very happy with the results.
Bass is tighter, seems faster and a litlle less blown, lower mids seem a tad more present, an impression that matches with the new curve.
But in my room it is almost impossible to obtain a good overall measurement of the complete speaker.
So i only can do (very) near field measurements of the individual speaker units.

Regarding Dirac/Mini dsp.
Was planning to go this way as well.
Would have been a lot easier.
But decided to stay analog.
Using a quality dac with the latest AK4499 converters to go from digital to analog, then convert everything back to digital-analog with the lower quality Dac's in the MiniDsp, was something i did not like.
Though maybe, this is not a problem at all, more a matter of preception.
Even had contact with MiniDsp.com, asking if they plan to introduce new gear in the near future with updated dac's.
Would have been nice to integrate the DSP with streaming and DAC in just one piece! (an updated SHD)
But this is not the case.
Still, MiniDsp is a fenomenal piece of equipment, that i am still thinking about to be honest.

 

[boomerang]

And the electric midrange output of the passive crossover.
Component values will be deliberately chosen to obtain higher output at low mids or compensate for accoustic properties of the FST unit of course.

 

[TH4DK]

Hello, im Torben from Denmark . I tried today to try my 2 pcs crown xls 2502 on the bottom of a set of b&w 801d4, with some big cambridge monoblocks on top and michi preamplifier. And the crown xls 2502 on the bottom Instead of cambridge monoblocks , is simply what these b&w speakers need, the sound went from thin and without body on the top and absolutely no bass, to a fantastic bass that can be adjusted on the crown gain control, and then the top became fantastic with body and fullness . Simply night and day experience

 

[GXAlan]

I haven’t heard the 801 D4’s myself, but it was recently reviewed and measured at Stereophile.

Bowers & Wilkins 801 D4 Signature loudspeaker

Back in August, I received an email from Editor Jim Austin. Subject line: "Want to do a big review?" He had my attention. Jim wrote that he had visited Bowers & Wilkins parent company Masimo Consumer in Carlsbad, California, for a demo of the brand-new B&W 801 D4 Signature and 805 D4 Signature...

www.stereophile.com

No surprises in its frequency response given what has been published, but this quote from B&W was interesting:

We don't measure or aim for a certain target in-room response. We're not tuning to a 'curve.' We develop the best engineering platform we can, and then we tune it so that we, as a group, can forget that we're listening to a pair of speakers and just get on with enjoying the listening experience. So, the target is this: Can you close your eyes and convince yourself, even for a split second, that the recording you're listening to is actually a real performer or group of performers in the room with you? If we manage to achieve that effect, we're happy. I've always said, no one buys loudspeakers to listen to loudspeakers. They buy them to listen to music."

I wonder if the FR irregularities measured nearfield are mitigated somehow by typical furniture or building materials or if the FR irregularity masking which occurs from stereo listening (instead of mono) mitigates certain irregularities more than others.

 

[blueone]

A rather tepid conclusion from JA for a $50K/pair speaker.

 

[thewas]

Horizontal directivity isn't too poor considering the size of the mid driver and lack of tweeter waveguide

hope that @John Atkinson will correct the wrong plot of the vertical one.

Bowers & Wilkins 801 D4 Signature loudspeaker Measurements

Sidebar 3: Measurements

www.stereophile.com

 

[Beave]

Horizontal directivity isn't too poor considering the size of the mid driver and lack of tweeter waveguide

hope that @John Atkinson will correct the wrong plot of the vertical one.

Bowers & Wilkins 801 D4 Signature loudspeaker Measurements

Sidebar 3: Measurements

www.stereophile.com

Keep in mind that this horizontal plot only goes from -45 to +45 degrees, whereas most of his horizontal plots go from -90 to +90 degrees. So if you're just comparing it at a glance with other speakers he has measured, it may look a lot better than it really is.